METHOD FOR DETERMINING THE POTENTIAL EFFICACY OF ANTICANCER TREATMENT
20200129566 ยท 2020-04-30
Assignee
- ASSISTANCE PUBLIQUE - HOPITAUX DE PARIS (Paris, FR)
- Institut National De La Recherche Agronomique (Paris, FR)
- UNIVERSITE PARIS-SUD (Orsay, FR)
- Institut Gustave Roussy (Villejuif, FR)
Inventors
Cpc classification
C12Q2600/106
CHEMISTRY; METALLURGY
A61K9/0065
HUMAN NECESSITIES
A61K9/0056
HUMAN NECESSITIES
G01N2800/52
PHYSICS
International classification
Abstract
Embodiments of the present disclosure relate to methods for ex vivo determining whether a patient with metastatic melanoma is likely to benefit from a treatment with an anti CTLA-4 molecule, preferably ipilimumab, by analyzing the gut microbiota in a fecal sample from said patient.
Claims
1-56. (canceled)
57. A composition comprising one or more purified bacterial strains with 16S rRNA sequences having at least 97% sequence identity with bacterial strains of species selected from the group consisting of Lachnospiraceae butyrate producing bacterium, Bacteroides ovatus, Ruminococcaceae clostridiales bacterium, Blautia obeum, Fusicatenibacter saccharivorans, Roseburia inulinivorans, Gemmiger formicilis, and Faecalibacterium prausnitzii.
58. The composition according to claim 57 comprising one or more purified bacterial strains of species selected from the group consisting of Lachnospiraceae butyrate producing bacterium, Bacteroides ovatus, Ruminococcaceae clostridiales bacterium, Blautia obeum, Fusicatenibacter saccharivorans, Roseburia inulinivorans, Gemmiger formicilis, and Faecalibacterium prausnitzii.
59. The composition according to claim 57 wherein the purified bacterial strains have 16S rRNA sequences having at least 97%, at least 98%, or at least 99% sequence identity.
60. The composition according to claim 57 wherein the composition comprises two or more, three or more, four or more, five or more, six or more, seven or more, eight or more, nine or more, or ten or more bacterial strains.
61. The composition according to claim 57 wherein at least 10%, at least 20%, at least 30%, at least 40%, at least 50%, at least 60%, at least 70%, at least 80%, at least 90%, or at least 100% of the bacterial strains belong to the Firmicutes phylum.
62. The composition according to claim 57 wherein less than 100%, less than 90%, less than 80%, less than 70%, less than 60%, less than 50%, less than 40%, less than 30%, less than 20%, or less than 10%, of the bacterial strains belong to the genus Bacteroides.
63. The composition according to claim 57 wherein the composition does not include bacterial strains of the genus Bacteroides.
64. The composition according to claim 57 wherein the composition does not include bacterial strains of the species Bacteoides fragilis or Bacteoides thetaiotamicron.
65. The composition according to claim 57 wherein the composition does not include bacterial strain Faecalibacterium prausnitzii A2-165.
66. The composition according to claim 57 wherein the bacterial strains are lyophilized.
67. The composition according to claim 57 wherein the composition further comprises an immune checkpoint inhibitor.
68. The composition according to claim 67 wherein the immune checkpoint inhibitor is a PD-1 inhibitor, PD-L1 inhibitor, or CTLA-4 inhibitor.
69. A pharmaceutical composition comprising the composition according to claim 57 and a pharmaceutically acceptable carrier.
70. The pharmaceutical composition according to claim 69, wherein the pharmaceutical composition is formulated for delivery to the intestine.
71. The pharmaceutical composition according to claim 69, wherein the pharmaceutical composition is in the form of a capsule.
72. The pharmaceutical composition according to claim 71, wherein the pharmaceutical composition is formulated for oral administration.
73. The pharmaceutical composition according to claim 69, wherein the pharmaceutical composition comprises a pH sensitive composition comprising one or more enteric polymers.
74. A method of treating cancer comprising administering a pharmaceutically effective amount of the pharmaceutical composition of claim 69 to treat the cancer in the subject.
75. The method of claim 74 wherein the cancer is melanoma.
76. The method of claim 74 further comprising determining if the subject has developed colitis.
Description
FIGURES LEGENDS
[0174]
[0175] A. Principal component analysis representation of patient distribution based on bacterial genera composition between baseline visit (V.sub.1) and colitis. Seven patients had fecal samples collected at both baseline and time of colitis onset. Monte-Carlo simulated p-value=0.0059. Component 1 explains 13.59% of variance; component 2 explains 11.45% of variance.
[0176] B. Relative abundance of dominant (>1% of total reads) gut microbial genera significantly reduced during colitis (V.sub.tox) as compared to baseline (V.sub.1). LACHN: Lachnospiracea incertae sedis, RUM: Ruminococcus, BLAU: Blautia, CL_IV: Clostridium IV, EUB: Eubacterium, Unc_LACH: unclassified Lachnospiraceae, PSEUD: Pseudoflavonifractor (paired t-test p<0.05 for all genera).
[0177] C. Proportions as percent of total reads of significantly impacted bacterial isolates during colitis (V.sub.tox) as compared to baseline (V.sub.1). Only significant (paired t-test p<0.05) data are presented.
[0178]
[0179] A. Principal component analysis representation of patients' distribution based on bacterial genera composition at baseline (V.sub.1) depending on the benefit of ipilimumab treatment; LT Benefit: long-term benefit (in green) vs Poor Benefit (in dark red). Component 1 explains 12.86% of variance; component 2 explains 8.67% of variance. Monte-Carlo simulated p-value=0.00899.
[0180] B. Boxplot of the percentages of 4 dominant (>1% of total reads) genera differentially represented between both groups, i.e. Bacteroides, Faecalibacterium, Clostridium XIVa and Gemmiger ; LT_Benefit: long-term benefit vs Poor Benefit; *:p<0.05; **:p<0.001.
[0181] C. Inter-class principal component analysis representation of patient distribution based on bacterial genera composition at baseline (V.sub.1) depending on overall survival time following ipilimumab treatment; GS0_6: overall survival ranging from 0 to 6 months (n=2), GS6_9: overall survival ranging from 6 to 9 months (n=5), GS9_12: overall survival ranging from 9 to 12 months (n=4), GS12_18: overall survival ranging from 12 to 18 months (n=7), GSsup18: overall survival superior to 18 months (n=8). Monte-Carlo simulated p-value=0.01098.
[0182] D. Percentages of 3 specific OTUs highlighted as biomarkers at baseline (V.sub.1) of overall survival duration greater than 18 months. Each patient's microbiota are presented in the graph. GS0_6: overall survival ranging from 0 to 6 months (n=2), GS6_9: overall survival ranging from 6 to 9 months (n=5), GS9_12: overall survival ranging from 9 to 12 months (n=4), GS12_18: overall survival ranging from 12 to 18 months (n=7), GSsup18: overall survival greater than 18 months (n=8).
[0183]
[0184] A. Inter-class principal component analysis representation of patient's stratification into 3 statistically robust clusters (A, B and C) at baseline (V.sub.i). Monte-Carlo simulated p-value=0.00099.
[0185] B. Bacterial genera discriminating the 3 different clusters at baseline were assessed with a Random Forest analysis, and confirmed with a Wilcoxon test. Percentage of reads for each of these 6 genera is represented for each cluster.
[0186] C. Baseline gut microbiota composition and overall survival (OS). Kaplan-Meier survival curves of patients classified into two groups according to clusters, Cluster A versus Cluster B.
[0187] D. Baseline gut microbiota composition and progression free survival. Kaplan-Meier curves of patients classified into two groups according to clusters, Cluster A versus Cluster B. P values are indicated on each graph.
[0188]
[0189] A. Boxplot of relative abundance of the 2 dominant phyla Firmicutes and Bacteroidetes at baseline between patients prone to or resistant to ipilimumab-induced colitis. A Wilcoxon test has been applied to assess significance and p-values are indicated on the graph.
[0190] B. Colitis cumulative incidence (Gray's test) of patients classified into two groups according to clusters, Cluster A versus Cluster B. P values are indicated on the graph.
[0191] C. OTUs predictive of colitis development during ipilimumab treatment. LEfSe uses Linear Discriminant Analysis (LDA) to estimate the effect size of each differentially abundant feature (i.e. bacterial OTU). On the left panel, OTUs in dark grey are biomarkers of colitis development (Colitis). OTUs light grey are biomarkers of the absence of colitis development (No_Colitis). Best biomarkers show the highest absolute LDA score with a minimal threshold of 3.5 (i.e. OTU denovo592 for the No_Colitis group and OTU denovo3795 for the Colitis group). Right panel indicates the taxonomic affiliation of each of these OTUs. Sim.: similarity between the OTU read and the first assigned isolate 16S sequence, assessed by the RDP Sab_score. The RDP S_ab score is a percentage of shared 7-mers between two sequences.
[0192]
[0193] A-C. Baseline (V.sub.1) percentages and absolute numbers of fresh whole blood CD4.sup.+ T cells, Treg cells (% CD4.sup.+CD3.sup.+CD25.sup.+Foxp3.sup.+ within CD4.sup.+CD3.sup.+ cells) ; 4.sup.+7.sup.+ among CD4.sup.+ T and CD8.sup.+ T cells, serum concentrations of IL-6, IL-8 and sCD25 were analyzed and compared between patients belonging to Cluster A and patients belonging to Cluster B (A); long term clinical benefit (LT benefit) and poor clinical benefit (Poor benefit) (B); patients with colitis (Colitis) and patients without colitis (No colitis) (C). Each dot represents one patient. P values are indicated on each graph; ns means not significant; Mann Whitney tests were used.
[0194] (D) The expression of ICOS on CD4+21 T cells was monitored in fresh whole blood before ipilimumab treatment (V.sub.1) and after one or two injections of ipilimumab (V.sub.2-3). Serum concentration of sCD25 was monitored prior to ipilimumab treatment (V.sub.1) and after one or two injections of ipilimumab (V.sub.2-3). Percentage of conventional CD4+ 24 T cells (Tconv) was defined as % CD3+CD4+ T excluding Treg cells. Graphs depict specific changes in immune activation over the course of ipilimumab treatment in patients belonging to Cluster A (white circles) and patients belonging to Cluster B (black circles). Each dot represents one patient. 1 P values are indicated on each graph. Mann Whitney (A-C) and Wilcoxon matched-pairs signed rank (D) tests were used.
[0195]
[0196]
[0197]
[0198]
[0199]
[0200]
[0201]
EXAMPLES
I. Patients, Materials and Methods
[0202] Patients
[0203] Patients with MM treated with ipilimumab were prospectively enrolled at the Gustave Roussy Cancer Campus between March 2013 and December 2014. Patients were informed of the study and consented to participate. This study was approved by the Kremlin Bicetre Hospital Ethics Committee (GOLD study: SC12-018; ID-RCB-2012-A01496-37) and all procedures were performed in accordance with the Declaration of Helsinki. Patients had a pre-specified clinical workup; feces and blood were collected at baseline (V.sub.1), prior to each ipilimumab infusion (V.sub.2, V.sub.3, V.sub.4), at the end of treatment (ie, 3 weeks after the last infusion; V.sub.5) and, if present, at the time of colitis occurrence (VTox). Ipilimumab was administered intravenously every 3 weeks at a dose of 3 or 10 mg/Kg and could be continued after V.sub.4, at a maintenance dose of one infusion every 12 weeks, in patients whose disease was controlled (response or stable disease). When immune-mediated enterocolitis (grade III) was suspected, patients were referred to the Gastroenterology Department of Bicetre Hospital. The diagnosis of ipilimumab-induced colitis was made in patients who had endoscopic signs of inflammation and no other cause of colitis, such as ischemia and infection (stool tests for bacterial pathogens and Clostridium difficile toxin had to be negative).
[0204] Response to ipilimumab was assessed by several criteria. Long-term clinical benefit was defined by response decrease of tumour burden50% relative to baseline, according to immune related response criteria [15, 16]) or stable disease (decrease of tumour burden of less than 50% but with less than 25% increase relative to nadir) for more than 6 months. Patients with poor benefit were defined as patients with a lack of long term benefit, i.e. with progression-free survival of less than 6 months (with immune related progressive disease defined as a confirmed increase in tumor burden 25% relative to nadir) [15, 16]. All responses were confirmed by a subsequent assessment no less than 4 weeks from the date first documented.
[0205] Bacterial Composition Assessment by High-Throughput Sequencing
[0206] Fecal samples were collected anaerobically at baseline (V.sub.1), prior to each ipilimumab infusion (V.sub.2, V.sub.3, V.sub.4), at the end of the treatment (V.sub.5) and at the time of colitis (V.sub.Tox) and were kept at 80 C. until analysis. Total DNA was extracted from fecal sample aliquots (150mg), as previously described using both physical and chemical lysis [17]. Culture-independent 16S rRNA gene sequencing was performed on 83 fecal samples (n=26 patients; Table S3). Both 454 pyrosequencing (Life Sciences, a Roche company, Branford, Conn., USA) and MiSeq (Illumina Inc, San Diego, Calif., USA) technologies were applied on the V.sub.3-V.sub.4 16S rRNA gene region. Gut bacterial composition was determined prospectively without knowledge of colitis or of clinical benefit, which was determined at the end of the study.
[0207] More precisely, for both pyrosequencing (454) and MiSeq sequencing, the V.sub.3-V.sub.4 region of the 16S rRNA gene was amplified with the following primers: V.sub.3F <<TACGGRAGGCAGCAG>> (V.sub.3F bac339F modified with R instead of G) (Wilson K H, et al. J Clin Microbiol. 1990) and V.sub.4R <<GGACTACCAGGGTATCTAAT>>bac806R. For 454, applied quality filters were minimum length=250 bp, maximum length=600 bp, minimum quality threshold=25, maximum number of homopolymers=6. For MiSeq, the filters were minimum length=200 bp and minimum quality threshold=20. Applying these filters, an average of 1639 and 7377 reads/samples were kept for further analysis respectively with 454 and MiSeq technology. High quality reads were pooled, checked for chimeras, and grouped into Operational Taxonomic Units (OTUs) based on a 97% similarity threshold with uclust software from QIIME. Finally 462,312 high-quality reads (average 4,623 reads/sample) were analyzed and grouped into an average of 349 Operational Taxonomic Units (OTUs) per sample (97% similarity threshold; OTUs or phylotypes) based on uclust software from QIIME [1]. Estimates of phylotypes richness and diversity were calculated using both Shannon and Simpson indices on the rarefied OTU table (n=1,000 reads). Singletons were removed and phylogenetic affiliation of each OTU was done by using ribosomal database project taxonomy [2] and performed from phylum to species level.
[0208] The applied sequencing technology is indicated for each sample (n=83) in the Table S3, as is the number of samples used for comparisons within the different part of the study. Statistical analyses were applied to decipher the influence of ipilimumab treatment, colitis development, and other clinical factors on the microbiota and to highlight a combination of microbial groups and immunological defects significantly involved in colitis development in MM patients.
[0209] Statistical testing for significant differences between all combinations of two groups was conducted using either paired Student or the Mann-Whitney test. The statistical 1 language R was used for data visualization and to perform abundance-based principal component analysis (PCA) and inter-class PCA associated with Monte-Carlo rank testing on the bacterial genera.
[0210] The Random Forest classifier was applied to relative abundance data of distributed bacterial genera and OTUs to assess the variable importance of microbial components in the classification of each patient to a certain phenotype [3].
[0211] The random forest machine learning analyses were applied to identify bacterial genera level that differentiates fecal community composition between groups. The purpose of a classifier such as random forest is to learn a function that maps a set of input values or predictors (here, relative to genera abundances) to a discrete output value (here, relative to clinical groups).
[0212] Random forest is a powerful classifier that can use nonlinear relationships and complex dependencies between taxa. The degree of the success of the method is its ability to classify unseen samples correctly, estimated by training it on a subset of samples, and using it to categorize the remaining samples [3]. The cross-validation error is compared with the baseline error that would be achieved by always predicting the most common category. Random forest analysis was performed for each comparison on 5000 rarefied versions of the data.
[0213] Additionally, random forest assigns an importance score to each genus by estimating the increase in error caused by eliminating that genus from the set of predictors. It also reports the Gini importance of a variable, which is computed as the sum of the Gini impurity decrease of every node in the forest for which that variable was used for splitting. Moreover, the LEfSe (linear discriminant analysis effect size) algorithm [4] was applied for high dimensional discovery of biomarkers that discriminate between patients that will have colitis but who may also respond (or not) to ipilimumab. The algorithm is freely available at http:huttenhower.sph.harvard.edu/galaxy/. Finally, patients were statistically stratified based on their dominant gut microbial composition (genus level). Clustering of patients was performed with the k-means clustering algorithm and Calinski-Harabasz index 1 to select the optimal number of clusters as previously described [5]. The k-means clustering algorithm implemented in R package was used. Interclass PCA of genera composition with clusters as instrumental variable was also assessed, based on a Monte Carlo test with 1000 replicates.
[0214] Lastly, random forest analysis was also performed to identify bacterial genera that differentiated fecal community composition between the three identified clusters.
[0215] Fresh Whole Blood Immune Monitoring and Soluble Immune Markers
[0216] Blood samples were collected at baseline (V.sub.1), prior to each ipilimumab infusion (V.sub.2, V.sub.3, V.sub.4), at the end of treatment (V.sub.5) and at the time of colitis (V.sub.Tox). Phenotyping was performed on fresh whole blood or on fresh peripheral blood mononuclear cells isolated by Ficoll density gradient and frozen for later analyses (see supplementary materials). All serum samples were stored at 80 C. until further analysis of soluble immune markers of inflammation (IL-6, IL-8, IP-10, MCP-1, TNF, sCD25) and bacterial translocation (sCD14). Details of immune monitoring can be found in supplementary materials. We monitored memory T cells and ICOS induction on T cells because previous studies had demonstrated that both could be related to clinical responses [18]. Regulatory T cells play a crucial role in the maintenance of immune homeostasis in the gut and are supposed to be a major target for anti-CTLA-4 treatment [19]. The heterodimer 47 plays a crucial role in the intestinal homing of T cells; we monitored 4.sup.+7.sup.+ T cells to gain some insight in the mechanism of ipilimumab-induced colitis. We monitored soluble immune markers of inflammation and bacterial translocation as a complement to microbiota composition. Indeed, microbiota composition affects gut barrier function, local as well as systemic immunity [20-25] [26]
[0217] Statistical Analyses
[0218] A formal sample-size calculation was not performed for this pioneering study. Associations between microbiota dominant profile and immunological parameters were assessed with the Spearman correlation coefficient and a two-sided Wilcoxon test. No adjustment for multiple comparisons was made because of the exploratory component of the analyses. Fisher's exact test was used to examine the association between microbiota clusters and immune parameters; microbiota clusters and colitis, as well as microbiota clusters and clinical responses.
[0219] Overall survival (OS) and progression-free survival (PFS) according to intestinal microbiota (Cluster A vs. Cluster B) and according to the number of conventional CD4.sup.30 T cells (low number of conventional CD4 vs. high number of conventional CD4) were estimated using the Kaplan-Meier method and compared using the log-rank test.
[0220] When considering the occurrence of toxicity, progression and death lead to informative censoring, because when patients are censored at progression or death, their risk of toxicity could differ from that of patients who are not censored. Competing risk analysis is a method to deal with competing events, in order to avoid selection bias induced by informative censoring.
[0221] Therefore, toxicity cumulative incidence function (CIF) was estimated through a competing risk analysis in which toxicity was the event of interest, death or progression was the competing event and patients alive without progression or toxicity at their last follow-up were censored from the date of last follow-up. The toxicity cumulative incidence functions in the different gut microbiota clusters and for the different conventional CD4 levels were estimated using the SAS macro %CIF and compared using the stratified Gray's test [27, 28]. These analyses were performed using SAS software, Version 9.4 (SAS Institute Inc., Cary, N.C., USA).
[0222] II. Results
[0223] Clinical Characteristics of the Patients
[0224] Fifty-five patients were included in the study and followed-up for at least 6 months. Among them, 13 patients did not receive ipilimumab, while four patients received only a single infusion of ipilimumab and died soon thereafter due to their melanoma; they were therefore excluded from further analyses. Among the 38 remaining patients, 26 provided fecal samples at baseline (i.e. before the first dose of ipilimumab) for microbiota analysis. Nine of these 26 patients (34.6%) had a long-term clinical benefit and seven (26.9%) developed an immune-mediated colitis.
[0225] Microbiota Composition is not Modified by Ipilimumab Treatment, except at the Time of Colitis
[0226] The gut microbiota was not significantly modified over the course of the ipilimumab treatment. Main bacterial phyla, Firmicutes and Bacteroidetes, remained stable over time. Even though bacterial genera proportions differed between the different time points, they were not significantly altered by ipilimumab treatment (Wilcoxon test p>0.05). Likewise, diversity was not altered by ipilimumab injections as assessed by both Shannon and Simpson indices. Altogether these data indicate that ipilimumab does not induce a gut microbial dysbiosis in patients with MM.
[0227] Seven patients provided fecal samples both at baseline (V.sub.1) and during colitis (Vtox). Immune-related colitis was associated with a shift in the gut microbial composition as highlighted on the principal component analysis (PCA;
[0228] Baseline Microbiota Composition is Associated with Subsequent Clinical Response and Immune-Related Colitis
[0229] At V.sub.1, inter-class PCA based on genera composition showed a significant clustering of patients depending on their clinical benefit in response to ipilimumab, either long-term clinical benefit or poor clinical benefit (Monte-Carlo test, p=0.00899;
[0230] When the microbiota composition of patients was considered at baseline independently from their clinical characteristics and without a priori hypotheses (k-means clustering algorithm and Calinski-Harabasz index), three groups of patients emerged that were defined as clusters: Cluster A, B and C. The inter-class PCA based on genera composition depending on belonging to any of these 3 clusters highlighted that this stratification was significant (Monte-Carlo test, p=0.00099;
[0231] At the phyla level, microbiota of patients prone to develop colitis was enriched in Firmicutes at baseline (Wilcoxon test, p=0.009), while high proportions of Bacteroidetes was observed in patients who did not develop colitis (p=0.011;
[0232] Baseline Immune Parameters Related to Microbiota Clustering, Clinical Benefit and Ipilimumab-Induced Colitis.
[0233] Twenty-six patients had both immune monitoring and microbiota composition analysis at V.sub.1. Patients from Cluster C could not be included in this analysis due to the low number of patients belonging to this cluster (n=4). Since patients who belonged to Cluster A (n=12) and Cluster B (n=10) had no differences in absolute counts of peripheral-blood CD4.sup.+ and CD8.sup.+ T cells (Table S5), the main T cell subpopulations were considered as a percentage. Patients who belonged to the Faecalibacterium-driven Cluster A had a low proportion of baseline regulatory T cells (Tregs), and had a significantly lower proportion of baseline 4.sup.+7.sup.+ CD4.sup.+ (p=0.0447) and 4.sup.+7.sup.+ (p=0.0344) T cells, compared to those who belonged to the Bacteroides-driven Cluster B (
[0234] Patients with long-term benefit tended to have a lower frequency of regulatory T cells (p=0.056) and a lower frequency of 4.sup.+7.sup.+ CD4.sup.+ (p=0.014) and 4.sup.+7.sup.+ CD8.sup.+ (p=0.0081) T cells, compared to those with poor clinical benefit (
[0235] Patients who developed ipilimumab-induced colitis tended to have significantly higher absolute numbers of CD4+ T cells (p=0.0529) and had significantly lower levels of IL-6, IL-8 and sCD25 at baseline compared to patients without colitis (
[0236] The Inducible T-cell COStimulator (ICOS) Molecule is Significantly Up-Regulated on CD4.sup.+ T Cells after Ipilimumab Treatment in Patients who belong to Faecalibacterium-Driven Cluster A
[0237] Since immune parameters were monitored longitudinally, specific changes in immune activation during the course of ipilimumab treatment were assessed in patients from Cluster A vs B. During ipilimumab treatment, Inducible T-cell COStimulator (ICOS) was significantly increased on conventional CD4.sup.+ T and Treg cells in patients who belonged to Cluster A (
[0238] III. Conclusion
[0239] This study shows that a distinct baseline gut microbiota composition is associated with an anti-cancer response and immune-mediated colitis in patients with metastatic melanoma treated with ipilimumab. Colonization by Firmicutes and more specifically by OTUs related to Faecalibacterium prausnitzii L2-6, butyrate producing bacterium L2-21 and Gemmiger formicilis ATCC 27749, is associated with both anti-cancer response and immune-related colitis. There is a higher representation of Bacteroidetes (mostly Bacteroides genus) in patients who have a poor anti-cancer response and who remain colitis-free.
[0240] In addition, in our dataset, Bacteroides fragilis and Bacteroides thetaiotaomicron were not abundantly represented at baseline. Overall, these bacterial species were more abundant in patients with poor clinical benefit who remained colitis-free.
[0241] We found that the Faecalibacterium-driven Cluster A and clinical benefit were associated with low baseline percentage of circulating 4.sup.+7.sup.+ T cells and CD4.sup.+ Tregs. In addition, high proportions of Faecalibacterium prausnitzii, low baseline percentage of circulating CD4.sup.+ Tregs and low baseline levels of systemic inflammatory proteins, such as IL-6 IL-8 and sCD25, were associated with subsequent colitis.
[0242] It has also been observed that patients who belong to the Faecalibacterium-driven Cluster A have a higher ICOS induction on CD4.sup.+ T cells and a higher increase in sCD25. Taken together, these observations show that gut microbiota composition at baseline can predict which patients will adequately respond to ipilimumab.
[0243] IV. Result of Metabolomics Studies
[0244] From the OTUs population described in the overall cohort, metabolomic profiles have been deduced in silico applying the PICRUSt algorithm combined with LEFSe.
[0245] A specific set of metabolic pathways (described as KEGG IDs) have been highlighted specifically associated with response to anti CTLA4 (see
[0246] High Throughput Metabolomics was Performed on the Fecal Samples at Baseline. Following metabolites extraction, sample were analysed by LC-HRMS as follows:
[0247] High-performance liquid chromatographic (HPLC) technique for separation of polar and hydrophilic compounds: aSequant ZIC-pHILIC (Merck, Darmstadt,Germany),
[0248] Gradient analysis for 30 minutes.
[0249] Q-Exactive mass spectrometry (Thermo Fisher Scientific),
[0250] High resolution analyses (70 000 FWHM) alternating positive and negative ionization modes.
[0251] For each metabolite (unique m/z & RT pair) identified, ion intensity (peak area) is reported for each sample. Semi-quantitative analysis can be performed comparing the relative intensity of a specific metabolite between samples and data are normalized as relative abundance (proportions) of each metabolite within one sample's profile.
[0252] Out of overall 256 detected metabolites, a combination of 14 metabolites can help decipher before aCTLA4 treatment whereas patients are more or less susceptible to answer adequately to the treatment. These 14 metabolites belong to 3 different sets. First, a combination of metabolites can be screened as biomarkers of response to aCTLA4. The LEfSe (linear discriminant analysis effect size) algorithm was applied for high dimensional discovery of biomarkers that discriminate between patients that will respond (or not) to ipilimumab. The algorithm is freely available at http://huttenhower.sph.harvard.edu/galaxy/.
[0253] These metabolic biomarkers are more often detected and in higher abundance in fecal samples of the clinical group they belong to.
[0254] Main metabolites that help discriminate between Responders and Non Responders are:
[0255] 1Trehalose-Sucrose-Isomaltose (p=0.0266)
[0256] 2Guanine (p=0.07523)
[0257] 3D-Raffinose (p=0.08581)
[0258] 43,4-Dihydroxyphenylacetic acid (p=0.0755)
[0259] which are more present in feces of Responders and
[0260] 5-12-Hydroxydodecanoic acid (p=0.006172)
[0261] which is more present in feces of Non responders (see
[0262] Individual repartition of these 5 metabolites in fecal samples of patients is represented in
[0263] A second set of metabolites (n=3), even though not specific of each samples from one clinical group, were significantly differentially represented between patients that responded and patients that did not respond to the aCTLA4 treatment. These metabolites are in Table VII below:
TABLE-US-00006 TABLE VII Wilcoxon pvalue Metabolite Response Delta 0.03128 1-Methylxanthine Non Responders 5x more 0.02251 Pipecolinic acid Non Responders 3x more 0.008023 2-Isopropylmalic acid Responders 4x more
[0264] Of note, isopropylmalic acid levels were also higher at baseline in patients that developed ipilimumab-associated colitis (1.6 more in patients with colitis)
[0265] Finally, a third set of metabolites (n=6) were selected that allowed stratification of the patients that were not properly classified with the previous 2 sets of metabolites. This 3.sup.rd set included:
[0266] 1Alpha D Amino adipic acid (higher levels in Non Responders)
[0267] 2Kynurenic acid (higher levels in Responders)
[0268] 3Methylimidazoleacetic acid (higher levels in Non Responders)
[0269] 43, 4 dihydroxyhydrocinnamic acid (higher levels in Non Responders)
[0270] 5Cotinine (higher levels in Non Responders)
[0271] 62 Methylnicotinamide (higher levels in Non Responders)
[0272] Taken together, these 14 metabolites allow the segregation between responders and non responders at baseline (PCA; component 1 explain 23.64% of variance, and component 2 explains 12.93% of variance; Monte-Carlo randtest test on the inter-class PCA simulated p-value=0.00789).
[0273] The ROC curve was built based on these 14 metabolites and the area under the curve showed good prediction value (AUC =0.8602941) (see
[0274] Distribution of the 14 metabolites within the 2 groups of patients is represented in the boxplots (y-axis=peak area) of
REFERENCES
[0275] 1. Robert C, Thomas L, Bondarenko I et al. Ipilimumab plus dacarbazine for previously untreated metastatic melanoma. N. Engl. J. Med. 2011; 364(26):2517-2526. [0276] 2. Hodi FS, O'Day S J, McDermott D F et al. Improved Survival with Ipilimumab in Patients with Metastatic Melanoma. N. Engl. J. Med. 2010; 363(8):711-723. [0277] 3. Schadendorf D, Hodi F S, Robert C et al. Pooled Analysis of Long-Term Survival Data From Phase II and Phase III Trials of Ipilimumab in Unresectable or Metastatic Melanoma. J. Clin. Oncol. 2015; 33(17):1889-1894. [0278] 4. Pardoll D M. The blockade of immune checkpoints in cancer immunotherapy. Nat. Rev. Cancer 2012; 12(4):252-264. [0279] 5. Postow M A, Callahan M K, Wolchok J D. Immune Checkpoint Blockade in Cancer
[0280] Therapy. J. Clin. Oncol. 2015; 33(17):1974-1982. [0281] 6. Weber J S, Kahler K C, Hauschild A. Management of Immune-Related Adverse Events and Kinetics of Response With Ipilimumab. J. Clin. Oncol. 2012; [0282] 30(21):2691-2697. [0283] 7. Beck K E, Blansfield J A, Tran KQ et al. Enterocolitis in Patients With Cancer After Antibody Blockade of Cytotoxic T-Lymphocyte-Associated Antigen 4. J. Clin. Oncol. 2006; 24(15):2283-2289.
[0284] 8. Gupta A, De Felice K M, Loftus E V, Khanna S. Systematic review: colitis associated with anti-CTLA-4 therapy. Aliment. Pharmacol. Ther. 2015; 42(4):406-417. [0285] 9. Marthey L, Mateus C, Mussini C et al. Cancer Immunotherapy with Anti-CTLA-4 Monoclonal Antibodies Induces an Inflammatory Bowel Disease. ECCOJC 2016; 10(4):395-401.
[0286] 10. Abraham C, Cho J H. Inflammatory Bowel Disease. N. Engl. J. Med. 2009;
[0287] 361(21):2066-2078. [0288] 11. Chassaing B, Darfeuille-Michaud A. The Commensal Microbiota and Enteropathogens in the Pathogenesis of Inflammatory Bowel Diseases. Gastroenterology 2011; 140(6):1720-1728.e3. [0289] 12. Lepage P, Hasler R, Spehlmann M E et al. Twin Study Indicates Loss of Interaction Between Microbiota and Mucosa of Patients With Ulcerative Colitis. Gastroenterology 2011; 141(1):227-236. [0290] 13. Vtizou M, Pitt J M, Daillre R et al. Anticancer immunotherapy by CTLA-4 blockade relies on the gut microbiota. Science 2015; 350(6264):1079-1084. [0291] 14. Sivan A, Corrales L, Hubert N et al. Commensal Bifidobacterium promotes antitumor immunity and facilitates anti-PD-L1 efficacy. Science 2015; 350(6264):1084-1089. [0292] 15. Wolchok J D, Hoos A, O'Day S et al. Guidelines for the evaluation of immune therapy activity in solid tumors: immune-related response criteria. Clin. Cancer Res. Off. J. Am. Assoc. Cancer Res. 2009; 15(23):7412-7420. [0293] 16. Snyder A, Makarov V, Merghoub T et al. Genetic basis for clinical response to CTLA-4 blockade in melanoma. N. Engl. J. Med. 2014; 371(23):2189-2199. [0294] 17. Seksik P. Alterations of the dominant faecal bacterial groups in patients with Crohn's disease of the colon. Gut 2003; 52(2):237-242. [0295] 18. Chen H, Liakou CI, Kamat A et al. Anti-CTLA-4 therapy results in higher CD4.sup.+ ICOShi T cell frequency and IFN-gamma levels in both nonmalignant and malignant prostate tissues. Proc. Natl. Acad. Sci. U. S. A. 2009; 106(8):2729-2734. [0296] 19. Read S, Greenwald R, Izcue A et al. Blockade of CTLA-4 on CD4.sup.+ CD25.sup.+ regulatory T cells abrogates their function in vivo. J. Immunol. Baltim. Md 1950 2006; 177(7):4376-4383. [0297] 20. Lukiw W J. Bacteroides fragilis Lipopolysaccharide and Inflammatory Signaling in Alzheimer's Disease. Front. Microbiol. 2016; 7:1544. [0298] 21. Guigoz Y, Dore J, Schiffrin EJ. The inflammatory status of old age can be nurtured from the intestinal environment. Curr. Opin. Clin. Nutr. Metab. Care 2008; 11(1):13-20. [0299] 22. Le Chatelier E, Nielsen T, Qin J et al. Richness of human gut microbiome correlates with metabolic markers. Nature 2013; 500(7464):541-546. [0300] 23. Sokol H, Pigneur B, Watterlot L et al. Faecalibacterium prausnitzii is an anti-inflammatory commensal bacterium identified by gut microbiota analysis of Crohn disease patients. Proc. Natl. Acad. Sci. U. S. A. 2008; 105(43):16731-16736. [0301] 24. Devillard E, McIntosh FM, Duncan S H, Wallace R J. Metabolism of linoleic acid by human gut bacteria: different routes for biosynthesis of conjugated linoleic acid. J. Bacteriol. 2007; 189(6):2566-2570. [0302] 25. Mathewson N D, Jenq R, Mathew A V et al. Gut microbiome-derived metabolites modulate intestinal epithelial cell damage and mitigate graft-versus-host disease. Nat. Immunol. 2016; 17(5):505-513. [0303] 26. Smith P M, Howitt M R, Panikov N et al. The microbial metabolites, short-chain fatty acids, regulate colonic Treg cell homeostasis. Science 2013; 341(6145):569-573. [0304] 27. Gray R J. A Class of $K$-Sample Tests for Comparing the Cumulative Incidence of a Competing Risk. Ann Stat. 1988; 16(3):1141-1154. [0305] 28. Lin G, So Y, Johnston G. Analysing survival data with competing risks using SAS* software. Proc. SAS Glob. Forum 2012 Conf. Cary N C SAS Inst. Inc 2012. [0306] 29. Riley J L, Blair P J, Musser J T et al. ICOS Costimulation Requires IL-2 and Can Be Prevented by CTLA-4 Engagement. J. Immunol. 2001; 166(8):4943-4948. [0307] 30. Lotze M T, Custer M C, Sharrow S O et al. In vivo administration of purified human interleukin-2 to patients with cancer: development of interleukin-2 receptor positive cells and circulating soluble interleukin-2 receptors following interleukin-2 administration. Cancer Res. 1987; 47(8):2188-2195. [0308] 31. Dubin K, Callahan M K, Ren B et al. Intestinal microbiome analyses identify melanoma patients at risk for checkpoint-blockade-induced colitis. Nat. Commun. 2016; 7:10391. [0309] 32. Maynard C L, Elson C O, Hatton R D, Weaver C T. Reciprocal interactions of the intestinal microbiota and immune system. Nature 2012; 489(7415):231-241. 33. Cebula A, Seweryn M, Rempala GA et al. Thymus-derived regulatory T cells contribute to tolerance to commensal microbiota. Nature 2013; 497(7448):258-262. [0310] 34. Png C W, Linden S K, Gilshenan K S et al. Mucolytic bacteria with increased prevalence in IBD mucosa augment in vitro utilization of mucin by other bacteria. Am. J. Gastroenterol. 2010; 105(11):2420-2428. [0311] 35. Hedin C R, McCarthy N E, Louis P et al. Altered intestinal microbiota and blood T cell phenotype are shared by patients with Crohn's disease and their unaffected siblings. Gut 2014; 63(10):1578-1586. [0312] 36. Romano E, Kusio-Kobialka M, Foukas P G et al. Ipilimumab-dependent cell-mediated cytotoxicity of regulatory T cells ex vivo by nonclassical monocytes in melanoma patients. Proc. Natl. Acad. Sci. U. S. A. 2015; 112(19):6140-6145. [0313] 37. Fu T, He Q, Sharma P. The ICOS/ICOSL Pathway Is Required for Optimal Antitumor Responses Mediated by Anti-CTLA-4 Therapy. Cancer Res. 2011; 71(16):5445-5454. [0314] 38. Liakou C I, Kamat A, Tang DN et al. CTLA-4 blockade increases IFN-producing CD4+ICOShi cells to shift the ratio of effector to regulatory T cells in cancer patients. Proc. Natl. Acad. Sci. 2008; 105(39):14987-14992. [0315] 39. Sim G C, Martin-Orozco N, Jin L et al. I L-2 therapy promotes suppressive ICOS+ Treg expansion in melanoma patients. J. Clin. Invest. 2014; 124(1):99-110. [0316] 40. Klatzmann D, Abbas A K. The promise of low-dose interleukin-2 therapy for autoimmune and inflammatory diseases. Nat. Rev. Immunol. 2015; 15(5):283-294. [0317] 41. Chaput N, Flament C, Locher C et al. Phase I clinical trial combining imatinib mesylate and IL-2: HLA-DR(+) NK cell levels correlate with disease outcome. Oncoimmunology 2013; 2(2):e23080. [0318] 42. Yu A, Snowhite I, Vendrame F et al. Selective IL-2 responsiveness of regulatory T cells through multiple intrinsic mechanisms supports the use of low-dose IL-2 therapy in type 1 diabetes. Diabetes 2015; 64(6):2172-2183. [0319] 43. Hodi F S, Chesney J, Pavlick A C et al. Combined nivolumab and ipilimumab versus ipilimumab alone in patients with advanced melanoma: 2-year overall survival outcomes in a multicentre, randomised, controlled, phase 2 trial. Lancet Oncol. 2016; 17(11):1558-1568. [0320] 44. Larkin J, Hodi F S, Wolchok J D. Combined Nivolumab and Ipilimumab or Monotherapy in Untreated Melanoma. N. Engl. J. Med. 2015; 373(13):1270-1271.